Once your AI is trained, you typically do not need to retrain it. However, retraining is required in the following specific situations:
When Retraining Is Required
Adding New Trading Sources
When: You have added a new data source to your AI configuration
Action: Train only the new source - do not retrain existing sources
Reason: Training existing sources again would create duplicate calculations and negatively impact AI performance
AI Configuration Changes
When: You have made adjustments to your AI settings or performed an AI reset
Action: Retrain all trading pairs completely
Examples: Changing profit percentage targets, modifying time-based parameters, or adjusting risk settings
Bot Migration
When: You have copied your AI configuration to use in a different bot with different markets
Action: Complete retraining for the new market environment
Reason: Different markets require specific training data to function optimally
Marketplace Strategy Updates
When: Your AI strategy includes a marketplace strategy with auto-update functionality, and that strategy receives updates
Action: Retrain your entire AI to incorporate the updated strategy
Critical: This step ensures your AI operates according to the latest rules and logic from the updated marketplace strategy
Consequence: Skipping this retraining may result in suboptimal AI performance
Strategy Update Integration
Process: When you retrain after marketplace strategy updates, the training process applies all modifications to every specific coin in your setup
Importance: Without retraining, marketplace strategy updates will not automatically apply to individual coins
Risk: Failing to retrain can cause inconsistencies in trading behavior and missed trading opportunities
Summary
Retraining is only necessary when your AI configuration changes, new sources are added, or when marketplace strategies are updated. Regular retraining without these triggers is unnecessary and may harm performance.